DE102016116630A1 - Method of operating an inverter and inverter - Google Patents

Abstract

A method for operating a transformerless inverter (1) having a first half-bridge (HB1) and a second half-bridge (HB2) arranged parallel to each other and to a DC link (DCL) between DC terminals (2, 3) of the inverter (1) are and whose bridge output (Br1, Br2) in each case via a filter choke (L1, L2) with one of the corresponding half-bridge (HB1, HB2) associated AC output (AC1, AC2) of the inverter (1) is connected, wherein the AC outputs (AC1, AC2) are connected to a network (5), and wherein between the AC outputs (AC1, AC2) is arranged a low-impedance coupled to the intermediate circuit (DCL) network (4) of filter capacitors, comprises operating the two half-bridges (HB1, HB2) of the inverter (1) with a unipolar clocking method and determining a value of a mains frequency leakage current at the DC terminals (2, 3) of the inverter (1). The method is characterized in that, when a limit value is exceeded by the leakage current value, the two half-bridges (HB1, HB2) of the inverter (1) are operated with a leakage-current-reducing clocking method in which the first half-bridge (HB1) generates an AC voltage at its assigned AC- Output (AC1), wherein an amplitude of the AC voltage is less than 50% of the amplitude of a voltage amplitude of the network (5) and the second half-bridge (HB2) a differential voltage between the mains voltage and the voltage provided by the first half-bridge (HB1) at the their assigned AC output (AC2). An inverter configured to carry out the method is also described.

Description

The invention relates to a method for operating an inverter, as well as an inverter.

Inverters are used in photovoltaic (PV) plants to convert DC power from solar modules into AC power for feeding them into a grid. Typically, the power co-generated by a plurality of solar modules is converted to AC by an inverter, with transformerless inverters allowing for particularly efficient conversion. In particular, single-phase feeding transformerless inverter with full bridge topology allow during operation usually no grounding of the solar modules, because the potential reference to ground potential changes with the grid frequency. This network-frequency potential change, in conjunction with design-related discharge capacitances on the solar modules, leads to leakage currents whose amplitude can assume increased values, for example in the case of rain. PV systems have differential current sensors for reasons of operational safety. Since the leakage current superimposed on the fault current of the system and contributes to a measured by the differential current sensors residual current, an increased leakage current can lead to an early triggering based on fault current measurements protection mechanisms, which automatically shuts off the PV system, although safe operation is still possible would.

It is therefore an object of the present invention to provide an operating method for a transformerless full bridge inverter, which counteracts an increase in the leakage current. Accordingly, an inverter is to be shown, which is less sensitive to the influence of a high leakage capacity in the determination of fault currents.

This object is achieved by the method of independent claim 1. The dependent method claims referring back to this claim show preferred embodiments of the invention. Furthermore, the object is achieved by an inverter of the independent device claim 11, whereby the dependent device claims referring back to this claim represent further embodiments of the invention.

One aspect of the invention relates to a method for operating a transformerless full-bridge inverter having a first half-bridge and a second half-bridge, which are arranged parallel to each other and to a DC link between DC terminals of the inverter and the bridge output in each case via a filter choke with one of the corresponding half-bridge associated with the AC output of the inverter, wherein the AC outputs are connected to a network, and wherein between the AC outputs a low-impedance coupled to the intermediate circuit network of filter capacitors is arranged. The method comprises the steps of operating the two half-bridges of the inverter with a unipolar first clocking method and determining a value of a line-frequency leakage current at the DC connections of the inverter. The method is characterized in that, when a limit value is exceeded by the leakage current value, the two half bridges of the inverter are operated with a second phase current derating current, in which the first half bridge provides an AC voltage at its associated AC output, wherein an amplitude of the AC voltage less than 50% of the amplitude of a voltage amplitude of the network and the second half-bridge provides a differential voltage between the mains voltage and the voltage provided by the first half-bridge voltage at its associated AC output. The unipolar clock method is characterized in that the inverter during a period whose duration corresponds to the duration of a line half-wave, is operated so that only one of the half-bridges is clocked.

Preferably, the alternating voltages provided by the half-bridges are sinusoidal voltages, but any other voltage curves are conceivable in which the combination of the voltage curves provided by the first half-bridge and the second half-bridge corresponds to the voltage profile of the mains voltage.

Advantageously, the amplitude of the AC voltage provided by the first half-bridge is selected as a function of the leakage current value. Preferably, the amplitude of the alternating voltage provided by the first half-bridge is chosen to be lower the higher the leakage current value is. This results in a corresponding reduction in the amplitude of the potential fluctuation at the terminals of a generator connected to the inverter, so that an increase in the Ableitstromwertes is counteracted or this increase is even completely avoided. In a specific embodiment, the amplitude is controlled so that a maximum value of the leakage current is not exceeded.

It is advantageous for the AC output associated with the first half bridge to be connected to a neutral conductor of the network and to the AC output assigned to the second half bridge with the phase conductor connect. This can be ensured, for example, by an installation instruction that one of the AC connections is to be connected to the neutral conductor. Alternatively, the assignment can be checked once or regularly by the inverter measuring the voltages at the AC outputs to earth and, based on this measurement, determining which of the two AC outputs is connected to the neutral of the network. The assignment of the half-bridges present in the inverter as the first half-bridge and the second half-bridge can then be selected accordingly or pointed to an incorrect connection. Alternatively, however, it is also conceivable that the inverter determines whether an increase or decrease of the leakage current value is achieved by the change from the first clock method to the second clock method and in the case that the leakage current value increases, the assignment of the half bridges of the inverter as a first half-bridge and replace the second half-bridge.

In addition to the change of the clock method, the inverter can continue to increase when exceeding the limit value by the leakage current value applied to the DC terminals DC voltage by driving an input-side DC / DC converter. The DC / DC converter can in this case act as a boost converter or as a buck converter and, in particular, keep the connected solar modules despite changing the DC voltage at the DC terminals in the MPP (maximum power point). Due to the increased DC voltage at the DC terminals, a higher partial voltage can be set with the second half-bridge, so that the first half-bridge only has to set a smaller proportion of the mains voltage and thus less leakage current is caused.

In a preferred embodiment of the method according to the invention, when the limit value is exceeded, one of the half bridges is operated in a controlled manner by the leakage current value and the other of the half bridges is operated in a controlled manner. Preferably, in this case, the first half-bridge is operated controlled. A controlled operation can be realized, for example, by using a predetermined clock pattern or a predetermined curve of the duty cycle for driving the half-bridge without taking into account a voltage present at the bridge output or a current flowing there. In this case, the regulated half-bridge sets the desired mains current.

In an advantageous embodiment of the method, the first half-bridge and the second half-bridge are operated synchronously with each other, ie with the same clock frequency or with a fixed time reference between switching times of the first and the second. In this case, as switching times, e.g. the switch-on or the switch-off or the midpoint of the duty cycle of two switches of the two half-bridges are selected. However, it is also conceivable to operate the two half-bridges independently of each other and even with different clock frequencies.

Since operation of the inverter in the second clocking method results in higher thermal loading of the switching elements and the reactors, it is advantageous to limit maximum inverter power to a lower value during operation with the second clocking method than with the first clocking method. A stronger limitation of the maximum inverter power in the second cycle method thus prevents overloading of the inverter.

Taking into account the higher converter losses of the inverter in the second cycle method, it is advantageous to select an amplitude for the provided by the first half-bridge AC voltage when exceeding the limit by the leakage current value, which is at most 30% of the amplitude of the mains voltage to a reduction of the leakage current to achieve that compensate for the higher converter losses of the inverter in the second cycle. In particular, it is advantageous to change the amplitude of the AC voltage provided by the first half-bridge in stages. It is advantageous to provide at least two stages in the second cycle method. Alternatively, it is also conceivable to specify an amplitude for the alternating voltage provided by the second half-bridge when the limit value is exceeded, in particular to select it as large as the applied DC voltage at the DC connections allows, for example half of this DC voltage or a slightly reduced one Value to choose. As the DC voltage at the DC terminals changes, the amplitude of the AC voltage provided by the second half-bridge can also be continuously adjusted. The amplitude of the AC voltage provided by the first half-bridge is adjusted accordingly to provide in total the amplitude of the mains voltage.

In order to keep the period in which the inverter must be operated to limit the leakage current with the less efficient second clocking method as short as possible, the inverter continuously or repeatedly determines the current leakage current value during operation with the second cycle method and compares it with a further limit value which is preferably determined as a function of the currently used amplitude value of the AC voltage provided by the first half-bridge. If the current leakage current value falls below this further limit value, this is a character that can be switched back to the first cycle without exceeding the limit again. The inverter then switches back to the first cycle method either immediately or after a preset time in which the leakage current value does not exceed the further limit value. In particular, when the amplitude of the alternating voltage provided by the respective half-bridges is selected as a function of the DC voltage applied to the DC terminals, the further limit value can be determined as a function of this DC voltage.

In the event that the second clocking method governs a fixed leakage current value, the first clocking method would return to when the amplitude of the AC voltage provided by the first half-bridge exceeds an amplitude limit.

Another aspect of the invention relates to a transformerless inverter, which is adapted for operation with the method described above or its embodiments. Here, the network of filter capacitors between the AC terminals may comprise a series connection of two filter capacitors, the center of which is connected to one of the DC terminals or to a center of a DC link designed as a shared intermediate circuit. Furthermore, the filter chokes of the inverter according to the invention are not magnetically coupled together.

The two half-bridges together form, optionally with additional components of the inverter bridge, in particular additional switches, an H4, an H5, an H6, a H6Q or a HERIC topology. An overview of the mentioned topologies can be found, for example, in the article "From H4, H5 to H6 - Standardization of Full-Bridge Single-Phase Photovoltaic Inverter Topologies without Ground Leakage Current Issue" by Jianhua Wang et al., Published in 2012 IEEE Energy Conversion Congress and Exposure (ECCE) , or to the H6Q topology in the Scriptures EP 2 237 404 A1 , During this operation, the additional switches of the inverter bridge are permanently held in a switching state within a half-cycle during operation with the second, leakage-current-reducing clock method, so that the switches of the two half-bridges can provide the desired current voltage value by high-frequency clocking. Thus, in the nomenclature of Wang's article, the additional switch S5 in the H5 topology is permanently closed (see 2a in the article by Wang), the switches S5 and S6 in the HERIC topology permanently open (see 2e ), or in the H6 topology, the switches of the half-bridge halves facing the DC connections, as seen from the bridge output, are permanently closed with two switches connected in series (S1 and S3 in FIG 2f , S3 and S4 in 2g , S3 in 4b , S4 in 4c and 4d , S1 in 4e ).

In one embodiment, the inverter according to the invention can have only one current sensor for determining the AC output current, which is arranged in particular at the AC output assigned to the second half bridge. If, during operation with the second clock method, one of the half bridges is controlled and one of the half bridges is operated in a controlled manner, the current sensor is preferably arranged at the AC output, which is assigned to the controlled half bridge.

In addition to the inverter bridge with the first and the second half-bridge, an inverter according to the invention can furthermore have a DC / DC converter, in particular a step-up converter, whose output is connected to the DC connections. In this embodiment, the DC / DC converter can be used to provide a higher DC voltage at the DC terminals during operation with the second clocking method than during operation with the first clocking method.

In another aspect, the operating method according to the application relates to a transformerless inverter, in which the inverter bridge in H4 topology with a first half-bridge and a second half-bridge, which are arranged parallel to each other and to a DC between DC terminals of the inverter is executed. The bridge output of both half bridges is in each case connected via a filter choke to an AC output of the inverter assigned to the corresponding half bridge, the AC outputs being connected to a network. Between the AC outputs in this case a niederimpedant coupled to the intermediate circuit network of filter capacitors is arranged. The method comprises operating the two half-bridges of the inverter with a first clock method and determining a value of a line-frequency leakage current at the DC terminals of the inverter and is further characterized in that when a limit value is exceeded by the leakage current value, the two half-bridges of the inverter with a discharge current reducing second clock method can be operated. In the first clocking method, the half-bridges are operated in common with a bipolar clock pattern, thereby providing opposing voltage waveforms at both AC outputs with an amplitude that is 50% of the mains voltage amplitude, respectively. In the leakage current reducing clocking method, the first half-bridge provides an AC voltage at its associated AC output, the less than 50% of a voltage amplitude of the network, preferably less than 30% of the mains voltage amplitude, and the second half-bridge at the associated AC output, a differential voltage between the mains voltage and the voltage provided by the first half-bridge voltage ready. This results in the first clock method, a fluctuation of the voltage of the DC terminals to earth potential of 50% of the mains voltage amplitude and the second clock this reduced fluctuation and thus a reduced value of the mains frequency portion of the leakage. The amplitude of the voltage provided by the first half-bridge is preferably selected as a function of the magnitude of the leakage current, in particular chosen such that the power-frequency proportion of the leakage current remains below a predetermined critical value. Preferably, this amplitude is changed stepwise if corresponding limit values of the power-frequency component of the leakage current are exceeded or fallen below. In this case, the amplitude is lowered when corresponding limit values are exceeded and increased when falling below it.

This alternative method of operation according to the application can be used, in particular, in conjunction with semiconductor switches which switch faster than silicon switches, for example of gallium nitride, whereby the switching and magnetization losses during the second clock method are not or not substantially increased compared to the first clock method.

In the following the invention and some of its variants are explained in more detail by means of figures, wherein

1 a schematic structure of an inverter according to the invention in H4 topology,

2 a further embodiment of an inverter according to the invention in H5 topology,

3 a time course of voltages during operation of the inverter according to the invention in a unipolar first clock method, and

4 shows a time course of voltages during operation of the inverter according to the invention in a leakage current reducing second clock method.

The in 1 shown inverters 1 has DC connections 2 . 3 on, in which a voltage source, not shown, in particular a PV generator, may be connected. Between the DC connections 2 . 3 a DC link DCL is arranged, and parallel to the DC link DCL a first half-bridge HB1 and a second half-bridge HB2. The first half-bridge HB1 can be formed from two series-connected semiconductor switches T3, T4. The center of the semiconductor switch is led out as the first bridge output Br1 from the first half-bridge HB1. Analogously, the second half-bridge HB2 can be formed from two series-connected semiconductor switches T1, T2 whose center is led out as the second bridge output Br2 from the second half-bridge HB2. The semiconductor switches may have an intrinsic or a separate antiparallel freewheeling diode.

A first filter choke L1 connects the first bridge output Br1 to a first AC output AC1, a second filter choke L2 connects the second bridge output Br2 to a second AC output AC2. Between the two AC outputs AC1, AC2 is a network 4 arranged by filter capacitors, which together with the filter chokes L1, L2 form an AC line filter. The network 4 is formed here by a series connection of two filter capacitors whose center is connected via a low-impedance connection 6 to the DC connection 3 connected. The low-impedance connection 6 Here is a direct connection, which is also conceivable to provide in the connection further, at the line frequency and at the switching frequency of the bridges having a low impedance components. Alternative to the DC connection 3 can be the center of the network 4 as well to the DC connection 2 or the center MP of a shared in this case DC link DCL be connected so that the respective potentials are coupled together. The first AC output AC1 may be connected to a neutral N of a network and the second AC output AC2 is connected to a phase conductor L of this network. It is conceivable that the line filter further filter components, in particular further filter chokes between the network 4 and the connected network.

Both half-bridges HB1, HB2 are controlled by controllers C1, C2 assigned to them. The controller C1 switches on the switches T3, T4 of the first half-bridge HB1 via pulse width modulation, and the controller C2 switches on the switches T1, T2 of the second half-bridge HB2. The current provided by the half-bridges is detected by current sensors CS arranged at the bridge outputs Br1, Br2. In this embodiment, by determining the difference between the measured values of both current sensors CS, a value of the leakage current can be determined. The mains frequency leakage current is the frequency component of the difference in the frequency of the connected network. Alternatively, the mains frequency leakage current can also be detected by additional sensors on the AC side or the DC side of the inverter 1 be determined in a previously known manner.

Another embodiment of an inverter according to the invention 1 show the 2 , At the DC connections 2 . 3 Here, a PV generator PV via a DC / DC converter BC, for example, a boost converter, connected. On the input side, such a DC / DC converter BC can also be provided in the other inverter topologies in order to convert the voltage of the PV generator PV to the voltage of the DC link DCL. Symbolic is still a leakage capacity 7 shown as the cause of a leakage current that connects the PV generator PV to ground GND. The inverter bridge is designed here as a so-called H5 bridge which, in addition to the transistors T1 to T4 of the two half bridges HB1, HB2, also has a further transistor T5 which connects the upper connection point of the half bridges HB1, HB2 to the DC connection 2 combines. The intermediate circuit DCL is here also designed as a divided intermediate circuit with a center MP. The output side is the network 4 the AC filter in addition to a series connection of two filter capacitors between the terminals L, N, to which the network 5 is connected, yet equipped with another capacitor, which is located directly between the two power terminals L, N. The midpoint of the series connection of the filter capacitors of the network 4 is here directly connected to the midpoint MP of the DCL DCL. In the embodiment shown here, only a single current sensor CS in the connection lines to the network 5 provided, which is also possible in the other conceivable embodiments.

As is well known, the additional transistor T5 in the H5 topology serves to connect the connected PV generator PV during the freewheeling phases from the connected network 5 to disconnect electrically. In the context of this invention, however, he fulfills this function only in operation with the first, unipolar clocking method. During operation with the second clocking method, T5 remains permanently switched on, thus enabling independent operation of the two half bridges HB1, HB2 at the DC link DCL.

In order to explain the operation of the invention in more detail, is in 3 initially a temporal course of voltages shown, as for example at the H5 topology 2 arises when it is operated in the first cycle. The mains voltage U ₀ has the known sinusoidal curve with an amplitude Û ₀ . In the unipolar clock method, only one of the two half-bridges HB1, HB2 is clocked during a half-wave of the mains voltage. In connection with the potential-free freewheel, which arises in a _{topology-dependent manner} by the clocking of further switches, the half-bridges generate opposite sinusoids U _AC1MP , U _AC2MP with half the network amplitude Û ₀ / relative to the potential of the _{DC center point} MP at the AC terminals AC1, AC2. second These two sinusoids add up to the course of the mains voltage U ₀ . As the AC connector AC1 fixed to the neutral conductor of the network 5 is connected, the potential of the intermediate circuit midpoint MP compared to ground potential, which is equal assumed here to simplify the potential of the N-conductor varies also sinusoidal with an amplitude equal to half the power amplitude U _0/2. This leads to a power-frequency component of a leakage current via the discharge capacity 7 which is proportional to the amplitude of the variation of the generator potential. This proportion is added to a residual current flowing due to non-ideal isolation of the PV generator PV and can, in particular, be at a high value of the leakage capacitance 7 lead to an initiation of insulation monitoring, although the system is still in a sufficiently isolated state. Therefore, in order to reduce the power-frequency component of the leakage current, it is desirable to reduce the network-frequency amplitude of the variation of the generator potential so as to likewise reduce the corresponding network-frequency component of the leakage current.

Here is the second clock method is used according to this invention for carrying, which is able to reduce this amplitude of the variation of the potential generator to a value less than half the power amplitude u _0/2. The operation of this second clock method is using the in the 4 shown voltage curves and illustrated.

This shows 4 again the sinusoidal profile of the mains voltage U ₀ in comparison to the voltage _curves U _AC1MP , U _AC2MP at the respective AC terminals AC1, AC2 with respect to the potential of the _{DC center} point MP. The amplitude Û _{AC1MP of} the voltage _profile at the AC connection AC1 is less than the amplitude Û _{AC2MP of} the voltage _profile at the AC connection AC2. Advantageously, the amplitude Û _{AC1MP is} at most 30% of the mains voltage amplitude Û ₀ . Accordingly, the amplitude Û _{AC2MP is} at least 70% of the mains voltage amplitude Û ₀ . Since, as already described above, the AC terminal AC 1 with the neutral conductor of the network 5 is connected, the voltage _{waveform of the DC} mid-point MP varies only with the smaller amplitude Û _AC1MP and is correspondingly reduced compared to the resulting from the first _{clock method} described above amplitude. Similarly, the power-frequency component of the leakage current is reduced by the second clock method.

In order to achieve the asymmetric voltage curves at the AC terminals AC1 and AC2, the respective half bridges HB1, HB2 are regulated independently of one another on the voltage curves corresponding target values of the bridge voltage. As an alternative to a voltage regulation in the half-bridges HB1, HB2, only one half-bridge, preferably the first one, can also be used Half bridge HB1, voltage-controlled operation, while the second half-bridge is operated current-controlled, so that there is a desired network current. The voltage regulation can be effected by determining the deviation of a measured voltage at the respective AC connections AC1, AC2 from the predetermined target values and a change of the respective duty cycles corresponding to the deviation with which the half bridges HB1, HB2 are operated. However, it is also conceivable to operate one of the two half-bridges, preferably the first half-bridge HB1, with a predefined clock pattern which leads at least approximately to a voltage curve at the AC connection AC1 _having the desired amplitude Û _AC1MP . In this case, it is only necessary to operate one of the two half bridges regulated.

The operation of the first half-bridge HB1 can be carried out independently of the operation of the second half-bridge HB2, ie in particular also with a different frequency. In this case there is no temporal correlation between the switching times of the bridge switches of the two half-bridges. However, it is equally possible to operate both half-bridges with the same clock frequency and in particular synchronously with one another, for example by synchronizing the midpoints of the switch-on periods of the first and the second half-bridges.

To provide the asymmetrical voltage waveforms at the AC terminals AC 1 and AC 2, both half bridges must be constantly clocked, that is, during both half waves alternately both switches of the half bridge are closed, so that at the bridge output of this half bridge phase, both the positive and the negative voltage of the DC connections 2 . 3 provided. This is the cause of the above-described increased converter losses during operation with the second clocking method compared to operation with the first clocking method.

The minimum voltage U _{DCmin of} the generator required in the second, leakage-current-reducing clock method is determined by the amplitude Û _AC2MP at the second AC terminal AC2 and is twice this value. This value is higher than the required minimum voltage U _{DCmin of} the generator when the inverter 1 is operated with the first _{clock method} (U _DCmin = Û ₀ , see 3 ). Therefore, it is advantageous or even necessary to increase the generator voltage when changing from the first to the second clock method to avoid falling below this minimum voltage. The adaptation of the generator voltage can be done by a DC / DC converter BC upstream of the inverter.

LIST OF REFERENCE NUMBERS

1

 inverter

2, 3

 DC connection

4

 network

5

 network

6

 connection

7

 leakage capacitance

AC1, AC2

 AC output

Br1, Br2

 bridge output

T1-T5

 switch

HB1, HB2

 half bridge

C1, C2

 controller

DCL

 DC

BC

 DC / DC converter

PV

 PV generator

L1, L2

 filter inductor

CS

 current sensor

GND

 ground

N

 neutral

L

 phase conductor

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• EP 2237404 A1 [0017]

Cited non-patent literature

• "From H4, H5 to H6 - Standardization of Full-Bridge Single-Phase Photovoltaic Inverter Topologies without Ground Leakage Current Issue" by Jianhua Wang et al., Published in 2012 IEEE Energy Conversion Congress and Exposure (ECCE) [0017]

Claims (16)

Method for operating a transformerless inverter ( 1 ) having a first half-bridge (HB1) and a second half-bridge (HB2), which are connected in parallel to one another and to a DC link (DCL) between DC terminals (HB1). 2 . 3 ) of the inverter ( 1 ) and whose bridge output (Br1, Br2) in each case via a filter choke (L1, L2) with one of the corresponding half-bridge (HB1, HB2) associated AC output (AC1, AC2) of the inverter ( 1 ), the AC outputs (AC1, AC2) being connected to a network ( 5 ), and wherein between the AC outputs (AC1, AC2) a low-impedance coupled to the intermediate circuit (DCL) network ( 4 ) is arranged by filter capacitors, comprising the steps: - Operating the two half-bridges (HB1, HB2) of the inverter ( 1 ) with a unipolar clocking method, and - determining a value of a line-frequency leakage current at the DC connections ( 2 . 3 ) of the inverter ( 1 ), characterized in that, when a limit value is exceeded by the leakage current value, the two half-bridges (HB1, HB2) of the inverter ( 1 ) are operated with a leakage-current-reducing clock method, in which the first half-bridge (HB1) provides an AC voltage at its associated AC output (AC1), wherein an amplitude of the AC voltage less than 50% of the amplitude of a voltage amplitude of the network ( 5 ) and the second half-bridge (HB2) provides a differential voltage between the mains voltage and the voltage provided by the first half-bridge (HB1) at its associated AC output (AC2). The method of claim 1, wherein the alternating voltages provided by the half-bridges (HB1, HB2) are sinusoidal voltages, respectively. Method according to Claim 1 or 2, wherein the amplitude of the alternating voltage provided by the first half-bridge (HB1) is selected as a function of the leakage current value, in particular the amplitude with a higher leakage current value is chosen to be lower. Method according to one of the preceding claims, wherein the amplitude of the AC voltage provided by the second half-bridge (HB2) in dependence on a at the DC terminals ( 2 . 3 ) voltage is selected. Method according to one of the preceding claims, wherein the AC output (AC1) assigned to the first half-bridge (HB1) is connected to an N-conductor of the network ( 5 ) connected is. Method according to one of the preceding claims, wherein, when the limit value is exceeded by the leakage current value, one of the DC connections ( 2 . 3 ) applied DC voltage by driving an input-side DC / DC converter (BC) is increased. Method according to one of the preceding claims, wherein when the limit value is exceeded, one of the half bridges (HB1, HB2) is operated in a controlled manner by the leakage current value and the other of the half bridges (HB1, HB2) is operated in a regulated manner. Method according to one of the preceding claims, wherein the first half-bridge (HB1) and the second half-bridge (HB2) are operated synchronously with each other. Method according to one of the preceding claims, wherein the first half-bridge (HB1) and the second half-bridge (HB2) are operated independently of one another, in particular with different clock frequencies. Method according to one of the preceding claims, wherein a maximum inverter power during operation with the first clock method is limited to a higher value than the second clock method. Transformerless inverter ( 1 ) having a first half-bridge (HB1) and a second half-bridge (HB2), which are connected in parallel to one another and to a DC link (DCL) between DC terminals (HB1). 2 . 3 ) of the inverter ( 1 ) and whose bridge output (Br1, Br2) in each case via a filter choke (L1, L2) with one of the corresponding half-bridge (HB1, HB2) associated AC output (AC1, AC2) of the inverter ( 1 ), the AC outputs (AC1, AC2) being connected to a network ( 5 ), and wherein between the AC outputs (AC1, AC2) a low-impedance coupled to the intermediate circuit (DCL) network ( 4 ) is arranged by filter capacitors, arranged for operation with a method according to any one of the preceding claims. Inverter ( 1 ) according to claim 11, wherein the network ( 4 ) comprises a series circuit of two filter capacitors whose center is connected to one of the DC connections ( 2 . 3 ) or with a center of a designed as a shared intermediate circuit DCL (DCL) is connected. Inverter ( 1 ) according to claim 11 or 12, wherein the filter chokes (L1, L2) of the AC outputs (AC1, AC2) are not magnetically coupled. Inverter ( 1 ) according to one of claims 11 to 13, having one of the bridge topologies H4, H5, H6, H6Q and HERIC. Inverter ( 1 ) according to one of claims 11 to 14, wherein only one current sensor (CS) is provided for determining the AC output current, which is arranged in particular at the second half-bridge (HB2) associated AC output (AC1, AC2). Inverter ( 1 ) according to one of claims 11 to 15, further comprising a DC / DC converter (BC), in particular a step-up converter whose output is connected to the DC connections ( 2 . 3 ) connected is.

Images